The recent discussion of Bonnie Nardi's _"A small matter of programming"_[1] got me to take it off my bookshelf and reread it. Here are some notes from my reading, in case anybody's interested. Particularly @jonathoda who asked for a summary. http://akkartik.name/post/nardi [1] https://futureofcoding.slack.com/archives/C5T9GPWFL/p1559061487146900

I've used that reference of Field of Dreams too 🙂

Thanks @Kartik Agaram. Being lazy here, because the book is also sitting on my shelf, but how does Nardi define “task-specific”? Spreadsheets son’t seem very task-specific. They are very opinionated about the kinds of models they can build, but they have been applied to all sorts of user tasks.

I think "task-specific" here refers to just two characteristics: • Not general purpose. (Hence the "specific".) • Close to the tasks someone wants to do. You don't have to combine too many primitives to get to something useful. Though the line of "too many" is fuzzy and subjective, of course. In general, I think "easy" really just means "no high energy barriers". Not everything is nearby, but some things are nearby, and the rest can be reached by gradual climbing without hitting any insurmountable obstacles.